Yes.
Kick it off.
Let's do it.
Hey, Doug.
Hey, how are you?
Yeah, we'll kick things off. Good afternoon. I'm Kyle Joseph. I cover the specialty finance and fintech spaces here at Stephens. Pleased to have with me the CFO of Pagaya Technologies, Evangelos Perros.
Correct.
How'd I do?
Very good.
Good.
You can call me EP to make it simple with me.
We're going to go with that. For those that don't know, Pagaya provides AI/ML-based credit decisioning tools that integrate with various lending partners to analyze and approve customer credit. The company funds loans with a mix of ABS and forward flow agreements and is currently focused on the personal loan, Auto, and POS side of the world. With that, I'll turn it over to EP to introduce himself and the company. I'll lead questions, but I would encourage you guys to not be shy and make this as interactive as possible. Go ahead, EP.
Thanks for the invite. Thanks for having me. And thanks everyone for joining us today. I've been with Pagaya for about four years now. I joined back in 2021 before we went public. Before that, I was with Apollo Global Management for five years, and before that with JPMorgan Chase. Maybe just to very high level introduce Pagaya, think about Pagaya as a B2B business that offers a white label solution to other lending institutions, partners, banks, fintechs, Auto captives, BNPL, or POS credit providers. We integrate with them on one side of our network and then create the pipes to directly feed institutional investors with these types of assets: personal loan, Auto, and POS.
We have 155 or so institutional investors on the other side, a mix of alternative asset managers and insurance companies, sovereign funds, everything you can think of, probably the largest ABS issuer in the personal loan space. If I had to highlight two or three key differentiators for Pagaya, one, it's a B2B business. We don't have a brand. We don't work with consumers directly. Everything we do is on behalf of our lending partners. We have 31 lending partners on the other side, companies like U.S. Bank, SoFi, Ally, Klarna on the POS. We integrate into their systems. Through that, basically, we make fees on both sides of the network. 80% today is actually coming from the lending partners as part of the use of our technology. That's at a high level just what Pagaya does.
Happy to go into more details, obviously, as we go through the discussion.
Great. You guys reported your third quarter, was it last week? Yeah, last Monday. Do you want to just walk us through some high-level kind of trends? I know you guys increased your guidance for 2025.
Yeah. Obviously, it has been a very strong year. You are basically seeing the results of a lot of the things that we did last year where we were 100% focused on getting the company to be GAAP net income positive and cash flow positive. In the last quarter, if I had to highlight a few things, one is it was the third quarter of GAAP net income, straight quarter of GAAP net income positive, run rating at closer to, when you analyze it, $110 million of net income. Credit performance is stable and steady and in line with expectations. The business has a very unique operating leverage. You look at the combination of qualitative consumer lender in many ways, but with the financial profile of a technology company, we do not have any customer acquisition cost or none of that marketing spend. It comes through the financials.
On top of that, we recently said and announced that on our pipeline of lending partners, eight of them now are actually in the onboarding phase and effectively planting the seed for growth into the later part of 2026 and 2027. We can talk about a little bit of dynamics there. I think what's more exciting probably to think about is you look at the numbers, application flow that we get through our lending partners grew quarter-over-quarter by 12%. Translation, the product strategy that we had in place is working. Volume actually increased quarter-over-quarter by 6%. Very prudent, very stable, consistent underwriting since the beginning of the year. We had positioned the company since the beginning of the year to anticipate protracted volatility and uncertainty. That 6% volume growth translated to a 10% fee revenue growth. We continue to monetize this volume at higher rates.
More importantly, translated to a 24% growth in adjusted EBITDA and, more importantly, 35% in net income. This is a great quarter when you actually look at the numbers. Not every quarter is going to be like this, but it is a good quarter where everything is coming together in the financials to really prove out what we are trying to do in our strategy.
Great. With that, let's talk about the topic du jour. Obviously, equity markets seem to think that the consumer's weakened recently. Can you walk us through, given all your various products, various consumers, what you're seeing in terms of the strength of the underlying consumer and your credit performance?
Yeah. I'll separate a little bit the macro and the credit. I would say on the macro, obviously, there has been a lot of focus from investors, and we're monitoring that, obviously. We always do, irrespective of the noise in the marketplace. We haven't seen any deterioration. A lot of people are talking about the tariffs, inflation, all of these things. We haven't seen yet the impact. If we do, obviously, we'll react. I would highlight that you may have seen a lot of other industry players, which I think most of them still talk about a healthy consumer. You may have seen a little bit of more volatility, like overshooting earlier in the year, now some sort of, call it, pullback. We haven't done that, and that's obviously causing a little bit of that stress in many ways.
From our perspective, and when you actually think about the consumer, we are effectively underwriting a very tight goal of 670 on the personal loan, $115,000 of income. So even the ones that may get impacted a little bit from macro inflation, what have you, that's more on the lower income side. We haven't seen any impact from a macro perspective in the data. And we get the benefit of working with 31 different lending partners from banks, Hifi, all the way to subprime lending providers. So we have the and across multiple asset classes, personal loans, Auto, and POS. The benefit is actually quite helpful. On the credit, credit performance, steady and in line with expectations. We haven't seen any change. Again, you always have to keep in mind what's the starting point or point of reference.
We actually started the year very prudently in terms of the underwriting, our conversion ratio, how much of the application volume that we see that would turn into funded volume. It's stable at 1%. Has been very stable for many quarters now. The biggest validation we get on the credit is from third-party investors. Everybody's talking about Auto in Q3, and in the same quarter, we announced a new forward flow with Castlelake on the Auto side. We sold certificates in our recent ABS to One William Street. That's the biggest validation. To take this a little bit to the next level and give actually some of the data, you look at personal loan on POS, pretty much in line with expectation and no real change in the data. In Auto, you did see some elevated delinquencies on the 60-day +.
You need to take into consideration when you're solving for returns on performance, you actually look at the recovery rates. They're substantially higher than what they were last year, as an example. Roll rates, how many of the 30% become 60%, 60 days- 90 days, and so on and so forth, much better than both 2024 and 2023. It is something obviously we're closely monitoring, as everybody is doing, but we haven't seen any change in the performance. Again, as I said, the biggest validator is demand from third-party investors in our production.
Yeah. Auto, in particular, seems to be getting picked on right now. I am glad you touched on that. In this macro environment, how would you classify your appetite for growth? Then kind of secondarily, how is your lending partner's appetite for growth? Are you seeing more loans? Is part of the application volume a function of your lending partners turning more folks out?
On the lending part, when you repeat to see our lending partners, by default, you have marketing dollars. When you want to push for growth, you spend more of those marketing dollars to get more application flow into your funnel and translate that into volume. There is going to be generally more variation, more volatility. People who know consumer lending and direct-to-consumer type of brands, you'll see that happens. For us, it is much more stable just to begin with because when they open up credit, we get less of the application flow. When they tighten credit, we get more of the application flow, and we adjust accordingly. It is less volatile on our end. What I would say is you had a mix of things. You had people who opened up earlier. You had some others being a little more stable. Some of them actually tightening for different reasons.
It's a mixed bag, but generally still very healthy. Again, from our perspective, the way we grow is this is, call it, on a same-store basis. In our minds, strategically, that's noise. Ultimately, the way we grow is by adding more partners to the network. Remember, it's quite unique and different business story. There was some overcorrection in the last few quarters, people talking about, "So what happened?" and a little bit of trickle. There is some noise in the market, and obviously, people are taking a different view on the stance. Again, from our perspective, when we think about our growth, to your question, you'll never see us grow 30%, 40%, 50% in a single quarter. We don't even have that lever to begin with. I don't have marketing dollars to spend. It's a B2B business.
As our partners grow, we'll grow with them, but it's a fully diversified portfolio and partners. Even if one grows by 50%, you're not going to see us grow by that much. We are growing with our partners and obviously product-led strategy. Generally speaking, the way you should think about us is like a 20% consistent sort of growth on an annual basis. That's how to think about it when you think about Pagaya.
Yeah. On that 20% growth, can you kind of walk us through how much of that is adding new partners? How much is leveraging existing partners? How should we think about new products, other products, and everything there?
Yeah. So the pillars of growth, I guess, for Pagaya is obviously existing and newer partners. Let's start with the existing partners. Within the existing partners, first, we get the benefit of their growth. When they're growing, all else being equal, we're getting more growth as well. Klarna is a good example, is a partner of ours. They're growing by 30% in the U.S. You have new deals, new merchants that they're adding to the network. As a result, we're getting that application flow as well. This is just pure organic growth in line with our partners. The second one is partners who are what we call enterprise-scale partners. So a partner who starts out, let's say, personal loans and they want to expand into Auto, we get that benefit as well, and we can partner with them.
We have a couple of examples in our current portfolio of sort of cross-selling across different products. The third one, and more exciting for us because it is something we have been focusing a lot in the last, call it, year or so, is what we call product-led growth. About 60% of our partners today are in one product. By product, I mean Declined Monetization, Affiliates, Prescreen, and things like that. These are different products for us to get more access to more of their application flow. And 60%, as I said, are now in the Declined Monetization. There is significant opportunity there to introduce some of these new products, and we see that already working out with some of the more established mature products. Lastly, the partners that we own, it takes a lot of time to ramp up a new lending partner.
It can be anything from six months, 9 months on the low end all the way to 12 months, 18 months. It does not suddenly open up. We do not open up the entire flow, either because of process issues or, more importantly, for risk. We need to make sure we understand. The model needs to adjust to the data that is getting from that flow from that specific partner. It takes time. When you think about these things, these are how to build up or stack up the growth potential for our business, and it is all organic growth. Obviously, the other one is new partners.
We said eight partners moved from the pipeline into the onboarding phase, but you should always think about that as, again, don't expect, even if we turn on the switch with those new partners, let's say, even in the next quarter or so, you won't see them contribute much on volume until the later part of 2026, if not even in 2027. It just takes a lot of time. If you add all these up, you can see how many avenues of growth we have. Again, we're solving for a 20% year-over-year growth. Just something to keep in mind as well. Why 20%? Why not more? We don't have the we don't need to push for growth. We have the operating leverage. We have the highest unit economics in our history, a very diversified set of portfolio.
Any incremental growth which we could push for would come with incremental, obviously, credit risk. We do not want that. We want much more product-led type of growth. Over time, that is like a key differentiator, I think, for Pagaya. Again, because we have the operating leverage, we have already the cabinet-ing of profitability, and we want to deliver steady, profitable, sustainable growth.
Great. Thanks. As we think about your pipeline of eight partners, can you walk us through kind of the dynamics of adding a new partner and what that does to your margin, specifically the RL?
So generally.
Yeah. That acronym.
FRLPC, Fee Revenue Less Production Cost, which is basically the revenues that we earn from the network. FRLPC, I know it's a mouthful. In last quarter, it was 5%. For every $100 of volume, we earn $5 in fees, of which 80% of that comes from the lending partner side. The other piece is just admin fees, management fees that we earn from the investment vehicle. Pretty much all of it is from the lending partners. To your question, the interests are very much aligned. You start with a new partner, and you will start at very low fee rates. At the beginning, effectively, they keep, let's say, all the fees that you earn upfront.
As you mature that relationship and you prove the value proposition to them and you do more volume with that specific partner, you start, on a marginal basis, earning more of those fees, like splitting those fees more to our favor. When it comes to big enough volumes, again, these are tiered contractual arrangements, on a marginal basis, they're willing to share more of those fees with us than they will actually retain for that specific loan. You need to, if you ask why, think about the value proposition. We're coming. We're giving them a solution to approve more customers on their brand. They maintain their relationship. They are the servicer. They make service fees throughout the life of the loan, and all of that without any capital to work.
As we do more with them, in some cases, we do 30%-40% more volume than they actually do, they're willing to share most of the fees generated from, let's say, that specific loan. It takes time to ramp them up. We have said 4%-5% we're going to expect to earn generally on our volume. The less mature partners start very low and ramp up again based on the timeframe I laid out earlier, call it up to 12 montsh-18 months post turning on the switch where the flow starts coming. The more mature are at much higher levels. In a lot of cases, even higher than 6% overall. Think about it as a portfolio of mature and newer partners, and that's how the dynamics flow.
Yeah. That's very helpful. Lastly, in terms of asset classes, you guys are in Auto, POS, and personal. Is there anything else on the horizon you're thinking about?
So 2024 was a—I'll give you a little bit of a hint. 2024 was a year where we cleaned up, let's just call it. Improve unit economics, further improve operating leverage, capital efficiency, and we're laser-focused on these three asset classes. We continue to be because, obviously, there is so much organic growth on the table for us based on what I laid out in terms of pillars. At the end of the day, though, strategically, when you think about the long run, we're there to be a service provider to our partners.
If they want to expand to different asset classes outside of what we can offer, as long as they fit within our technology and underwriting, a little bit shorter duration type of assets, not like mortgages that have a seven-year duration and things like that, I think that's where you can see us continue to play and be next to them. Things that we have already, we're sort of looking on, again, not right now a priority given the ROI that we have in the existing asset classes, but home improvement is an interesting one for us. Healthcare POS is an interesting one. Credit cards as well, though we need to figure out a little bit the J-curve situation there. Things that are obviously in the consumer space have this sort of risk-reward type of characteristics that consumer loans, like personal loans, Auto, and POS have.
We're looking, obviously, for things that are big-time and have the similar structure like POS or BNPL or personal loan where you have multiple constituents that come into play, and there is a fee bucket that gets shared in every loan. Those are the most attractive opportunities for us, I guess, in the long run.
All right. On the funding side, we talked a little bit about equity markets viewing the consumer as weakening. How are the credit markets? Obviously, you guys were able to, are still able to complete Auto ABS and even sell the residuals, which to me signals a very healthy market. Equity markets may be telling us something different. Yeah, walk us through what you're seeing on the credit side.
Yeah. The secular trend, let's call it, of private credit that we'll all hear is still there. You still have alternative asset managers, call it, that have pockets both for the opportunistic credit as well as the, call it, yield type of credit that goes more into the insurance capital. They have a lot of big pockets there, and they continue to play. Overall, that trend, combined with the constructive macro environment of, generally speaking, over time, lower rates, you see significant demand from the credit side. People are looking to lock yields for longer periods of time. They still have capital to deploy. That is still there. You may have variations in spreads a little bit. You had it in April Liberation Day. You had 550 basis points spread widening. You'll have these dislocations, variations, but the trend is still strong.
You see what's happening in the marketplace, what equity markets are saying. Even look at the announcements today across multiple fintechs, how many forward flows even were even announced outside of Pagaya. That trend is there, and that's like a tailwind for the industry. There will be fluctuations. There will be winners and losers. I think scale makes a big difference. The diversification of different papers, different partners is going to be an important thing. We don't see any changes into that trend. There will be dislocations. There always are. There is no real change in the demand, and then you see it across. It's not just Pagaya. You see it across the spectrum of, especially, generally, consumer credit. People are looking to put dollars to work, institutional investors meaning, into these types of loans given the risk-reward profile that they have in the short duration.
We don't see any change to that. In our portfolio, from our perspective, a year ago, a little bit over a year ago, we were 100% funded through ABS. It was a very constructive effort to make sure we diversify the funding. Today, we're approximately 60//40 split, 60 being more like the ABS that we do, and then, call it, between structures where we hold the risk and the structures where somebody else holds, like forward flows or pass-throughs or ABS where we sell the certificates. That's where the current makes sense, and the goal is to be at that 50/50% mark. Diversification is always going to be a benefit to every player in the long run. I get the question, why not 100% forward flow? You don't want to be dependent on one single player or their performance.
You always want to have that diversification of funding.
Yeah. That's a good segue. As we think about, call it, your funding mix, how do we think about the economics of the forward flow versus the ABS? Is one more attractive? Obviously, it's probably fluid, and it depends on market conditions.
The surprise to people here, the economics, from a purely economics point of view, it's not that different. The way it shows up in our financials, in everyone's financials, is a little bit different. I'll walk through that. Generally speaking, obviously, forward flow is much more committed. It comes every month. Generally, if it's a, call it, let's call it $1.2 billion forward flow, it comes in a form of over 12-month terms. It's $100 million every month. You have a little bit more predictable, steady. You can plan with it and work with it. ABS, much more seamless, much more frictionless. You obviously take the risk of you have to go to market at that point. Now, for us, for those who don't know, we only do 100% pre-funding ABS, so we never really take liquidity risk, ever.
We always have, we go out, do an ABS, lock the cost of capital, and we always have this dry powder, which comes at a premium that we're willing to pay to not have liquidity risk. In terms of the economics, one, in the ABS, where with the sponsor, let's say, we'll have to have risk retention. In the forward flows, you won't have any risk retention, and it's not on the balance sheet, but it will come, let's call it, with a pricing concession, which will be reflected in our fees, like negative revenue, call it, or lower revenue. That's how a little bit the economics work.
If you actually look at the, today, for example, last three quarters or so four quarters, when you look at the risk retention as a result of this mix of forward flow and not in ABS, it's approximately 2%-2.5% on a net per dollar of volume. Compare that to the 5% FRLPC we earn, and that's how to think about the economics generally in our opinion.
All right. Yeah, in terms of risk retention, are you guys pretty happy with the 60/40 kind of funding mix you talked about, and where do you see risk retention trending over time?
I think 50/50, and there's no magic to the number. I keep on saying in a few meetings, it's 50/50. It's a very good mix where it accomplishes a few things. First, if you go well above, let's say, 50% on the forward flow, even for somebody like us, you need to be thoughtful about the allocation to the different vehicles, right? Think about it. We get every day, or every day, call it $35 million-$40 million, let's say, of loans that get funded, and we directly allocate them into the different mixes. Today, it's basically pro rata. If you have different mix that you need to adhere to, like that 50/50 allows you to be very flexible. Remember, we're a network. If we grow lending partners, we also need to grow the network on the other side.
The 60/40 that we are today is across all asset classes. We're slightly ahead of that on the personal loan because we started earlier. Auto, we just announced our first forward flow. POS is coming up. That is how we're at that 60/40 on the blended basis. 50/50, again, without real science behind the math, is a good mix.
Talk a little bit about the competitive environment. Obviously, Pagaya is a very unique model. What would prevent someone from replicating that model, and what would they need to do in order to do that?
Lose a lot of money in the process as we did. It's actually, we're quite unique. If you take Pagaya and we talk about B2B, look at even the peer set that analysts use, right? They use SoFi, LendingClub, Klarna, Affirm, Upstart. Three of those names that I just threw out are lending partners of ours. It's quite unique. There's nobody else in the marketplace that does what we do at scale. There's no other B2B that does this white label solution, which is quite unique. The barriers to entry are very high. It'll take you hundreds of millions of dollars to build the infrastructure. It will take somebody many years to build the data. In order to get the data, you need to get the partners. In order to get the partners, you need the data. It's very difficult.
Having the first mover advantage makes a huge difference. It doesn't mean that somebody else won't do it, but nobody does it today. It's going to take a lot of money and a lot of, call it, upfront infrastructure build-out that unlikely somebody can do or raise the money to go down that path at this scale and get the support from VC capital. There are players out there, very small ones that do more like a licensing fee, similar type of thing, but it's extremely small. It's more like, call it, used as a software to validate some sort of other work that the lender is doing more so than actually offer a solution to the problem. It's a lot of years, a lot of losses, a lot of infrastructure build-out. Again, even when we did it, there was nobody else.
For somebody to come in and replicate that can be done. It is going to be extremely challenging in our mind. That is a great position to be in, obviously.
Can you walk us through, as you're pitching a potential lending partner, maybe it varies if they're a bank or if they're a fintech, but the value add that you guys provide to them and how you're able to really convert them and get them in your funnel?
It's obviously a very long process, timing-wise. You need dedicated resources from both ends. Every partner is solving for something else. If you think about banks, there are banks who have automatic cut-off for, let's say, at 725. It's not that they underwrite the loan and they say no. They just don't even underwrite it. They just completely bypass it. We can come in and offer that solution to them under their brand. They're the originator on record, and we have all the infrastructure built out to adhere to their regulatory standards. That's why I was talking about barriers to entry. You have to really build that fair lending. That is just something that you have to address when you go and work with the banks. We have proven that out with U.S. Bank and SoFi and others.
The value proposition, though, for the banks is they do not want to lose that consumer who applies for a loan and would otherwise get a decline, and they take their money and go to a fintech player, as an example. They want to trap the deposits. They want to be able to cross-sell that to that consumer, other products. We can build their FICO score, so next time they refinance, they are ineligible. All of that is a good value proposition for the banks. Now that the regulatory environment, let's say, is, I do not want to say loosening up, but it is a little bit more constructive, they want to enter into POS type of credit and other aspects where we can really help them out. That is a little bit when you think about banks.
When you think about the allies of the world or Auto captives, all they care about, or primarily they care about, higher activation rate and satisfaction rate at the dealership level. They're competing with some of the other smaller fintech players. Less about the fees, much more they want to have multiple different offers and do it seamlessly to the consumer at the point of sale, meaning the dealership, and they just want to increase activation rate. That's the products that we have put in place, like Fast Pass, Dual Look, and things like that, which I'll have to talk about. POS lenders, all they care about at this point, given the growth that we see in BNPL, is growing the network of merchants, more merchants. How do they do that? They need to offer the solution.
They're very good, generally speaking, at the, call it, short duration products, traditional BNPL, buy now, pay later, four months type of thing. We can come in and help them with the longer duration, larger ticket items where it's interest-bearing and more simulates the personal loan type of structure and underwriting that we have in place. Everybody's solving for something else. We have a fintech that does subprime, and they do up to 640, and they want us for 640 and above FICO score. Everybody's now quite unique. We now have such a diversified set of partners and have been battle-tested in multiple different types of relationships where we can actually customize the solution to the lending partner. It just varies quite a bit.
Okay. We don't think about it as just a turn-down business. It can be a turn-up business.
Maybe to that, people think of Pagaya as like, okay, anything that people do not underwrite, it goes to Pagaya, and therefore it is higher risk, or why do you accept something that the others do not? And 60% of our partners today are with one product, and it usually starts with a Declined Monetization. Again, it is not pure second look. Second look is somebody underwrote it, said no, and then passed it over. Again, when U.S. Bank says automatic cut-off at 720, they do not even underwrite it. You need to keep in mind a little bit these types of dynamics. We have other products as well. I think that is the thing that people do not appreciate. We have Dual Look. Dual Look is we actually compete offers-wise with our partner, but obviously two offers come from the same partner. We do not really compete.
It's like the partner has two offers, one of their own and one of ours. They still come under the name of the partner, and the consumer now can choose. For example, one can be, I'll give you a good example, Ally. Ally will go out to a consumer and offer in the point of sale a $30,000 loan at 19% rate. We go through Ally and give $35,000 loan at 21% rate. We gave $5,000 more credit to that consumer at a higher rate, and the consumer can choose. That optionality is extremely helpful to the dealer. That's like a dual look, like side by side. We actually have first look products as well, or where we actually, for certain aspects of the flow, we're the first look. It's not a traditional second look, and the Declined Monetization is obviously the entry point.
That's something to keep in mind. This transition and getting more access to more flow to the pillars of growth I mentioned earlier, one of them is these new products that we're trying to do.
I got it. In terms of, do you get any pushback when you're pitching a lender? Or I guess phrased differently, why wouldn't a lender have Pagaya as an option?
We get pushback. The pushback is like, okay. We go in and pitch. We give them case studies, let's say, of the partners that we have worked with and the value that we have delivered. The business case, we have basically never really gotten a pushback. The business people love it. Obviously, why wouldn't they? When you go to a bank, when you go to the, it does take resources. In a bank, you have to go through risk. You have to go through cyber. You have to go through tech. You have to go through compliance and then go back up, back up, back up. By the way, what are you asking them to do? You're asking them to allow you as a fintech, in their minds, to get integrated in the loan origination system. It's not like that simple.
It's like somebody getting plugged in into somebody comes to you and says, "I'm going to plug in into your security system at home." You're worried. It just takes time, and it does take resources. For example, as I said, we have no customer acquisition costs, but incremental monetary investment to attract a new partner. It takes 100 people to work in an integration. It takes other people on the other side as well and some monetary sort of investment. We've never really gotten the no. It's just time to really educate them on the value proposition, explain to them how that works. At least where we are now in our evolution relative to two or three years ago, two years ago, we onboarded U.S. Bank, and then that opened up, obviously, the dialogue with all the banks.
Now we're at the point where we have multiple examples of partners in all of the products. And once they see the evolution and how much we can offer to them, we never really get the no for the answer, but it just takes time to really take them from that initial discussion all the way to something that's meaningful, obviously, for both sides.
Great. Thank you. And then so we, let's see. Your business is unique and some might say relatively complicated. I mean, just in terms of capital markets changes, walk us through how if ABS spreads widen, if AAA ABS spreads widen 25 basis points, how does that flow through your P&L or vice versa if they tighten?
Yeah. So the way to think about it is, by the way, whether ABS or, I mean, forward flow is a little bit different. But the way to think about it is think about every single ABS effectively balanced by itself. You have the assets, right? And they earn a certain return. And then you lock that cost of capital at the onset and say, "Okay, the cost of the cost of capital is X," X being 6%-7%, let's call it today. If spreads widen, effectively that 6%. Let's take an example. The assets return 9%, and the credit cost of debt is, call it, 7%. If spreads widen, it will cascade to all the tranches. For example, what happened in April or hasn't really happened today, effectively instead of 7%, it will be 7.5%. That's how to think about it. This by itself, right?
Therefore now you have a little bit less buffer. The return of the assets is, let's say, 8.5%-9.5%. The other one now came from 7%-7.5%. Effectively that closes the sort of buffer that you have and the expected return ultimately. How that translates to us as Pagaya, effectively that will bleed into the value of the certificate to the extent we hold it, the first loss tranche. That's how to think about it. Ultimately, the only way you'll see that is either if it's in a back, if the return changes on the backbook or something that we have already underwritten, it's in place, you'll see it more in the form of fair value adjustment and impairments.
Or in this new transaction, what you'll see is us, I'll give you an example, we may decide to say, "You know what? We're going to increase the threshold of approvals," as an example, and therefore not sold for 9%. On the asset side, we'll sell for 9.5%. We can very easily absorb this type of movement given our scale. If you're talking about something going up 500 basis points, you're talking about a completely situation. Again, even in that case, we'll always have dry powder in place from the previous deal because of the pre-funding or the forward flows. That's how to think about it where it actually cascades down to the P&L.
Got it. We're running up on time, so I'll probably combine two questions into one. It is 2025, so we'd be remiss if we didn't talk about AI and how it is involved with your business. From there, I think it's probably a good segue to talk about the scalability of the business. We talked a lot about the margin at the top, but in terms of the scale below that.
Let's start with the second one. The benefit of this business model, the B2B, is the infrastructure is already built out. There is no incremental investment required whatsoever to support even double the amount of volume. The infrastructure is there. Obviously, you'll have to spend a little bit more on cloud capacity. The infrastructure is built out. There is no need to invest. The expenses are completely uncorrelated with the incremental volume and top-line growth. That's why if you look at our P&L, it's much more of a technology-like P&L versus a consumer lender because we don't have this marketing cost. If you see us expand to other new asset classes, we may need to build a little bit more in getting the right team in place, things like that. Again, very little investment from here on because we have gone through all that pain already.
To your first question, I want to be clear. AI is not going to be, AI itself is not the one that really differentiates the underwriting in terms of its predictability of future losses. AI cannot predict. AI makes a very well data-driven model, machine learning. AI allows you to make a much more informed decision today based on what you know today. It cannot predict when interest rates are going to come down, when the next war in Ukraine is going to happen, when the COVID, hopefully not, is going to—it cannot predict that. It just makes a more informed decision in a much more efficient way today relative to manual processes. Because if you think about a credit box in a bank, let's say, or in a traditional model, it's very rigid and very sort of linear.
What the AI allows you to do is to actually identify pockets of money through constant cuts of the application of the data that we have to really understand the propensity and the ability for consumers, A, to pay over time and, more importantly, their sensitivity to pricing and therefore solve for the high returns between the different models. That is the benefit of AI. It cannot predict future losses. If you think that is what the AI does, you are up for a disappointment, just to be clear. That is how it compares to a manual process and why we can actually underwrite. Even us, by the way, if you ask us to underwrite a $500 FICO, it is tough. The model is not calibrated for these things. You just have to sort of have data.
It's just in our case, we have data from 31 different partners, multiple asset classes, very wide range from banks to non-banks to Auto to across the board. That allows you to actually make this more informed decision. Why we say there is a benefit to this relative to some of the other players? It's not because the model or the code is better. Anybody can write a code. The question is the data. If you have data across 31 different partners and you see something, think about it as a matrix, right? You have all the partners and multiple channels, and you have the different asset classes. If everything is red, turning something red, you get a signal. The level of confidence you have to react on something happening systemic, it's much stronger than anybody else's flow.
It's stronger even than a big bank who actually sees the flow that they are actually looking at, and that's the only thing they approve. It's bigger, obviously, than any other fintech that actually looks at their own flow. That's the benefit of the data advantage that we have of working with multiple different partners. That's a data advantage. The code is the code. AI is not here to, if there was AI that could actually predict the future in terms of losses, trust me, I wouldn't be here today. I would be printing money in my house or something. I want to be clear on that.
Okay, great. I think we're out of time. Anything you want to wrap up with? Or I apologize.
Open to questions.
Any questions from the audience? Thank you very much for your time.
No problem. Appreciate it.
Great to be with you. Thank you to everyone else who attended.
Thank you.
Appreciate it.
Thank you.